Merge pre/v2.1 into main

[dimitri-yatsenko]

Summary

Merge pre/v2.1 documentation updates into main.

Changes

Navigation

• Add missing nav entries for existing how-to guides (docs: Add missing nav entries for existing how-to guides #135)
  • manage-secrets.md → Setup section
  • object-storage-overview.md → Object Storage section
  • choose-storage-type.md → Object Storage section

Table Declaration Spec

• Remove ___ as alternative separator (only --- now accepted)
• Fix terminology: "secondary attributes" → "dependent attributes"
• Add section on hidden attributes (_-prefixed)
  • Behavior in heading, fetch, joins, restrictions
  • Access patterns and use cases

Query Algebra Spec

• Document dict restriction behavior
  • Unmatched keys silently ignored
  • Empty match returns all rows
  • Rationale and caution notes

Test plan

• Docs build without errors
• Navigation shows new entries
• Spec pages render correctly

🤖 Generated with Claude Code

[MilagrosMarin]

Thanks for merging the v2.1 updates! I verified the technical claims against datajoint-python pre/v2.1 source:

✅ Verified:

• ___ separator removal (only --- in pre/v2.1 declare.py:363)
• Hidden attributes behavior (heading.py:232, condition.py:376)
• Dict restriction unmatched keys (condition.py:375-387)
• Nav target files exist on pre/v2.1

Found a couple of items that need attention - see inline comments.

[dimitri-yatsenko]

Addressed Review Comments

Thanks @MilagrosMarin for the thorough review!

Comment 1 & 2: fetch() → to_dicts() / to_pandas()

Fixed in commit 79149c0. Updated both the behavior table and examples:

| to_dicts() / to_pandas() | Excluded unless explicitly projected |

# Visible attributes only (default)
results = MyTable.to_dicts()

# Explicitly include hidden attributes  
results = MyTable.proj('result', '_job_start_time').to_dicts()

# Or with fetch1 for single row
row = (MyTable & key).fetch1('result', '_job_start_time')

Comment 3: "dependent attributes" → "secondary attributes"

Fixed in commit 1812809. Standardized on "secondary attributes" throughout to match the Python API (heading.secondary_attributes). Removed the bridging note since we're now using the correct term consistently.

[MilagrosMarin]

Thank you, @dimitri-yatsenko! This PR looks good to me now.